README.md

---
title: Intelligent Web-Search RAG
emoji: 🌐
colorFrom: blue
colorTo: yellow
sdk: docker
app_port: 7860
pinned: false
---

🌐 Intelligent Web-Search RAG Chatbot
An advanced Retrieval-Augmented Generation (RAG) chatbot that combines the reasoning power of Llama 3.2 with real-time web search capabilities. Unlike standard AI, this agent can browse the live internet to answer questions about current events, world leaders, and breaking news.

🚀 Key Features
Real-Time Web Intelligence: Uses a custom-built DuckDuckGo search tool to fetch current facts beyond the model's training cutoff.

Conversational Memory: Remembers previous interactions using ConversationBufferMemory for a natural, multi-turn chat experience.

Intelligent Reasoning: Powered by the ReAct (Reason + Act) framework, allowing the AI to decide when to search the web and when to answer from memory.

Optimized for Performance: Implements smart result truncation to prevent context overflow and handles API rate limits gracefully.

Deployment Ready: Fully configured for local hosting or cloud deployment (Hugging Face Spaces, Docker).

🛠️ Tech Stack
Core Logic: LangChain (Agents, Tools, and Memory)

LLM: Meta Llama 3.2 3B Instruct via Hugging Face Endpoint

Search Engine: DuckDuckGo Search API (ddgs)

Web Framework: Flask (Python)

Environment: Python 3.10+ / Anaconda